A DETAILED EXAMINATION OF THE EFFECT OF WORKERS' COMPENSATION LAWS ON EMPLOYEE AND EMPLOYER BEHAVIOR

DEBORAH CLARY WELLAND, Purdue University

Abstract

This research explores the effect of increasing workers' compensation benefits on the resources allocated to accident prevention (safety production) by employees and employers. It is hypothesized that reducing the proportion of accident costs borne by employees will result in fewer accidents being prevented by them and/or more injuries being faked or exaggerated. Counter to this effect, the increase in costs borne by employers (and their insurance companies) will tend to increase their prevention and monitoring efforts. Prior research (verified by this study) has found that injury rates are positively correlated with benefit level suggesting a dominance of the employee effect. These hypotheses were tested using a 1976 microdata set which includes information on individual workers' compensation cases for seven states. A logit model was employed to estimate the relative probability of an employee-caused accident as a function of the benefit variables and personal characteristics of the employee. Because of a lack of a clear, usable definition of an employee-caused accident in the industrial safety literature, three variations of the dependent variable were used: employee-caused, employee-faked/exaggerated, and minor accidents. Within each of these categories several variations were used to ensure completeness and to study consistency of results. The sign of the coefficient of the independent variable measuring the income replacement ratio (weekly benefit/weekly wage) was of primary interest. If the employee effect dominates the coefficient is expected to be positive whereas domination of the employer effect produces a negative coefficient. The various forms of the equation were tested over a variety of industries in order to compare results. A majority of the significant coefficients on the income replacement variable were negative, indicating a dominance of the employer effect. However, no clear patterns of positive or negative signs by industry or dependent variable emerged. This study is a first pass at measuring the types of accidents which seem to be most likely to increase in relative frequency as benefits rise. Although the results are not conclusive, it opens the door for future empirical investigation in this area.

Degree

Ph.D.

Subject Area

Labor economics

COinS